path: root/drivers/md/dm-mpath.c

/*
 * Copyright (C) 2003 Sistina Software Limited.
 * Copyright (C) 2004-2005 Red Hat, Inc. All rights reserved.
 *
 * This file is released under the GPL.
 */

#include "dm.h"
#include "dm-path-selector.h"
#include "dm-hw-handler.h"
#include "dm-bio-list.h"
#include "dm-bio-record.h"

#include <linux/ctype.h>
#include <linux/init.h>
#include <linux/mempool.h>
#include <linux/module.h>
#include <linux/pagemap.h>
#include <linux/slab.h>
#include <linux/time.h>
#include <linux/workqueue.h>
#include <asm/atomic.h>

#define MESG_STR(x) x, sizeof(x)

/* Path properties */
struct pgpath {
	struct list_head list;

	struct priority_group *pg;	/* Owning PG */
	unsigned fail_count;		/* Cumulative failure count */

	struct path path;
};

#define path_to_pgpath(__pgp) container_of((__pgp), struct pgpath, path)

/*
 * Paths are grouped into Priority Groups and numbered from 1 upwards.
 * Each has a path selector which controls which path gets used.
 */
struct priority_group {
	struct list_head list;

	struct multipath *m;		/* Owning multipath instance */
	struct path_selector ps;

	unsigned pg_num;		/* Reference number */
	unsigned bypassed;		/* Temporarily bypass this PG? */

	unsigned nr_pgpaths;		/* Number of paths in PG */
	struct list_head pgpaths;
};

/* Multipath context */
struct multipath {
	struct list_head list;
	struct dm_target *ti;

	spinlock_t lock;

	struct hw_handler hw_handler;
	unsigned nr_priority_groups;
	struct list_head priority_groups;
	unsigned pg_init_required;	/* pg_init needs calling? */
	unsigned pg_init_in_progress;	/* Only one pg_init allowed at once */

	unsigned nr_valid_paths;	/* Total number of usable paths */
	struct pgpath *current_pgpath;
	struct priority_group *current_pg;
	struct priority_group *next_pg;	/* Switch to this PG if set */
	unsigned repeat_count;		/* I/Os left before calling PS again */

	unsigned queue_io;		/* Must we queue all I/O? */
	unsigned queue_if_no_path;	/* Queue I/O if last path fails? */
	unsigned saved_queue_if_no_path;/* Saved state during suspension */

	struct work_struct process_queued_ios;
	struct bio_list queued_ios;
	unsigned queue_size;

	struct work_struct trigger_event;

	/*
	 * We must use a mempool of mpath_io structs so that we
	 * can resubmit bios on error.
	 */
	mempool_t *mpio_pool;
};

/*
 * Context information attached to each bio we process.
 */
struct mpath_io {
	struct pgpath *pgpath;
	struct dm_bio_details details;
};

typedef int (*action_fn) (struct pgpath *pgpath);

#define MIN_IOS 256	/* Mempool size */

static kmem_cache_t *_mpio_cache;

struct workqueue_struct *kmultipathd;
static void process_queued_ios(void *data);
static void trigger_event(void *data);


/*-----------------------------------------------
 * Allocation routines
 *-----------------------------------------------*/

static struct pgpath *alloc_pgpath(void)
{
	struct pgpath *pgpath = kmalloc(sizeof(*pgpath), GFP_KERNEL);

	if (pgpath) {
		memset(pgpath, 0, sizeof(*pgpath));
		pgpath->path.is_active = 1;
	}

	return pgpath;
}

static inline void free_pgpath(struct pgpath *pgpath)
{
	kfree(pgpath);
}

static struct priority_group *alloc_priority_group(void)
{
	struct priority_group *pg;

	pg = kmalloc(sizeof(*pg), GFP_KERNEL);
	if (!pg)
		return NULL;

	memset(pg, 0, sizeof(*pg));
	INIT_LIST_HEAD(&pg->pgpaths);

	return pg;
}

static void free_pgpaths(struct list_head *pgpaths, struct dm_target *ti)
{
	struct pgpath *pgpath, *tmp;

	list_for_each_entry_safe(pgpath, tmp, pgpaths, list) {
		list_del(&pgpath->list);
		dm_put_device(ti, pgpath->path.dev);
		free_pgpath(pgpath);
	}
}

static void free_priority_group(struct priority_group *pg,
				struct dm_target *ti)
{
	struct path_selector *ps = &pg->ps;

	if (ps->type) {
		ps->type->destroy(ps);
		dm_put_path_selector(ps->type);
	}

	free_pgpaths(&pg->pgpaths, ti);
	kfree(pg);
}

static struct multipath *alloc_multipath(void)
{
	struct multipath *m;

	m = kmalloc(sizeof(*m), GFP_KERNEL);
	if (m) {
		memset(m, 0, sizeof(*m));
		INIT_LIST_HEAD(&m->priority_groups);
		spin_lock_init(&m->lock);
		m->queue_io = 1;
		INIT_WORK(&m->process_queued_ios, process_queued_ios, m);
		INIT_WORK(&m->trigger_event, trigger_event, m);
		m->mpio_pool